Initial Thoughts from a New User

[Sean]

I'm a fairly recent arrival to the PixInsight world, and thought I would share my initial thoughts, especially with other people who are thinking of taking the plunge. I saw Rogelio's presentation at AIC last year, which piqued my interest. I finally downloaded the Trial version a few months ago, and purchased PixInsight about a month ago. I'm an experienced Images Plus user (which I still really like), and also use CCDStack, RegisStar, Nebulosity, Photoshop, etc, etc. I'm a software developer by profession, so actually enjoy purchasing and using new software - weird huh?

To save the suspense, I really like PixInsight!

Positive things:
- The object orientation of the workflow seems very natural
- Platform independence is great - I do most of my image processing on a Mac now
- The look and feel of the software is very professional. Little things like using the Spacebar for panning, like in Photoshop. It's the little touches that make good software feel right.
- Extensibility - the opening of the API for script development is so powerful. Thanks to all who contributed to the Animation script!
- Pure power - I feel that I have complete control over calibration, registration, stacking, etc.  I started using mostly the defaults for each process, and am now venturing deeper and deeper as I gain a better understanding of how things work.
- Dynamic Background Extraction (DBE) - what a great utility.
- Wavelets - I'm just starting to understand the power of scale-wise image decomposition and processing. The more I learn, the more it seems as if this is the correct and natural way to process astro images.
- The User Community - thanks especially to Harry and his videos, and to Rogelio's unofficial guide

So, I'm a happy user and am looking forward to the release of V 1.7 with the capability to save and restore workspaces and processing environments.

I would encourage all those people who have looked at PI, but not yet tried it, to download the trial, watch Harry's videos, and start processing!

Sean

PS - I do have a few User Interface comments/questions which I will post separately in the appropriate forum.

[Juan Conejero]

Hi Sean,

Welcome to PixInsight Forum!

Thank you so much for your nice and encouraging words. I hope the more you know about PixInsight, the more you'll like it. You're a developer, so I'm sure that soon or later you'll get involved in PI development —developers are always hungry for more programming challenges, aren't we?

[Film Guy]

I've been visiting PI for quite awhile, and have decided to try the trial version.  Before I go any further, however, as my "name" implies - I shoot film.  I know PI has been specifically developed for astro imagers, and the newer versions are probably more specific to the digital realm.

My question would be - how well does it apply to film images?  I shoot mostly wide field nebulae - no galaxies.  I use mostly lenses for my Pentax 67 - and one has ED glass, others are "normal" lenses, so a bit of CA can be an issue on bright stars.

Secondly - I have CS4.  I realize it's limitations to an "average guy" who does not want to spend his lifetime learning the program.  I would really like to have a program that deals specifically with my requirements - tutorials to show me the way - and where I can develop a routine for processing. 

I'm sure this is all within the realm of PI.  I would just like to hear it from the guys who know - and many of you have shot film, and know what I'm talking about. 

Thank you.

[Carlos Milovic]

Hi "Film guy"

As you, many of us started with film too. And many of the tools in PixInsight are a legacy of those times. Even more, there are a few processes that were  done with film in mind. For example, see at the Divide process (now in the "Obsolete" category). It is aimed to apply flat field calibration frames done with film. Take a look at this old tutorial I worte: http://pixinsight.com/tutorials/obsolete-LE/filmflats-cmilovic/en01.html
The procedure has been automatized and enhanced in the process I mentioned earlier.

Also you'll find ACDNR, the successor of SGBNR, quite good to remove noise in your images.

[marekc]

Here's another reply for `Film Guy':

I've been playing around with film recently, and I will probably do a project with film this summer. (I'm basically interested in the extent to which a 35mm camera body can be used as an `emergency subsitute' for one's CCD camera.)

I recently shot some B&W film frames of the Lagoon Nebula, and then scanned the film at 4000 dpi with a Nikon Super Coolscan 5000(E?) film scanner. I used PI to register and stack them, to improve the SNR. I only stacked several 10-minute frames, so I wasn't able to completely eliminate the grain, but I could see a definite improvement in the SNR. (The standard deviation of the pixel values in a background area had gone down by a factor of `several', IIRC.)

I'm looking forward to experimenting with this some more. And thanks for the link to the flat-fielding tutorial, Carlos! I think I will also experiment with shooting actual flat-field frames on film, and seeing how that works, once I digitize them.

[Film Guy]

Hi "Film guy"

As you, many of us started with film too. And many of the tools in PixInsight are a legacy of those times. Even more, there are a few processes that were  done with film in mind. For example, see at the Divide process (now in the "Obsolete" category). It is aimed to apply flat field calibration frames done with film. Take a look at this old tutorial I worte: http://pixinsight.com/tutorials/obsolete-LE/filmflats-cmilovic/en01.html
The procedure has been automatized and enhanced in the process I mentioned earlier.

Also you'll find ACDNR, the successor of SGBNR, quite good to remove noise in your images.

Thank you Carlos.  I know you came from a film background as I have enjoyed your work for many years now.  The processes you talk about were my concerns, as the new technologies keep advancing the software to process the information.  As long as the basics of the procedures to develop the images still applies to film - after all - it has been digitized - then PI should be a valuable program.

That was my real concern.  That the new PI would apply to the "obsolete" called film yet still provide a real benefit over using PS for astro images?

I would also ask your opinion with regards to an experiment I am considering with film.  I love Ha, especially when combined with RGB images done with CCD.  I would like to push the envelop using E200 (I have a large stash in the freezer) where I would shoot 3 shorter exposures of 20 minutes for example as compared to a single 45 minute.  From one of the exposures, convert it to B&W and save as such.  Then stack and combine the B&W.  I feel this may give a different depth perspective of the object eg: M8 with the massive star cloud. 

I have a 4x5" camera that I would like to use for such a purpose using Fuji Astia 100 film, in combination with their B&W film Acros.  Using 150mm apo lens, it may show a massive area of the Milky Way without resorting to mosaic imaging?  I also have a roll film back for the camera, to allow me to shoot 120 format E200 film but in 6x9 size.

Carlos, I'm relying on your previous experience to ask for guidance with these projects.  I'm simply trying to do something different with film that has possibly not been done before.  If I can produce 2 or 3 incredible images per season, then I would be a happy man.

[Film Guy]

Here's another reply for `Film Guy':

I've been playing around with film recently, and I will probably do a project with film this summer. (I'm basically interested in the extent to which a 35mm camera body can be used as an `emergency subsitute' for one's CCD camera.)

I recently shot some B&W film frames of the Lagoon Nebula, and then scanned the film at 4000 dpi with a Nikon Super Coolscan 5000(E?) film scanner. I used PI to register and stack them, to improve the SNR. I only stacked several 10-minute frames, so I wasn't able to completely eliminate the grain, but I could see a definite improvement in the SNR. (The standard deviation of the pixel values in a background area had gone down by a factor of `several', IIRC.)

That's an interesting project.  I have a number of questions regarding the film used, as Ha response is low with most of them.  But also to invite you to visit the Film Forum on Cloudy Nights.  We could discuss the issues and with the experience of others, maybe help with this project and others you may consider.  It is a small dedicated friendly group.  Issues regarding your film grain may be related to the film type used and reciprocity.  Lots of B&W activity there with the Barnard objects and dark nebulae.
I'm looking forward to experimenting with this some more. And thanks for the link to the flat-fielding tutorial, Carlos! I think I will also experiment with shooting actual flat-field frames on film, and seeing how that works, once I digitize them.

[Carlos Milovic]

Yes, don't worry. Most of the processes will work with film images just fine. You may have some troubles with those processes aimed for linear data, such as all the deconvolutions. Anyway, with the use of masks, and forgiving the "theorical background", it may still be of good benefit for your images. Also you may find trouble with GREYCStoration, since it was created for digital camera's noise reduction, but you still have ACDNR and wavelets to do that job (and SCNR). So, to sumarize, just go ahead

Oh, btw, I believe that with the current image processing tools you'll be able to get better results than in the past. If it were not for the digital cameras, results from film would be getting better and better each day. Most of the techniques apply just the same.

I think that your project is worth a try. Combining the Ha data to the RGB ccd would not be an easy task, although. I would process both sets up to a certain point where they look pleasant to the eye, and then perform the combination. The luminance match will be the hardest part. You may try the color correction procedure I'm using with the HDRC process. I've tried that to fix saturation artegacts generated by deconvolutions, and it worked like a charm. I'm planning to write a standalone process with this procedure this week. With it, your combination should be easier.
Other than that, I may be a little concern about the different sampling size, and also by the different amount (and size) of the stars. That would be tricky too.
Anyway, it is a very nice experiment and would generate results very special. If you are capable of generate a new, fresh look of our well known objects, even when it is not  sharper, deeper, etc etc., then it will be worh all the effort. I'd love to see what happens. Keep us informed!


PS: Stacking film images does improve SNR a great deal. I used to combine up to a dozen exposures, and it was amazing how smooth it got, and how much details I was able to get.